Which component first in a loop?

[MrWhy]

So I've read things on both sides of the fence and the engineer in me belives the right answer to be that when possible the CPU should always come first (Flow from pump that is) and then the GPU in your loop. But I'm curious on how quantitative the advantage is in doing this, has anyone ever done their own tests on this subject? I'd love to hear your thoughts if you have thanks!

[Mi26]

I dont think there's a big difference

[MrWhy]

4 minutes ago, Mi26 said:

I dont think there's a big difference

Physically and mathmatically speaking there is, and it is affected by a bunch of things not limited to and including the emmisivity of the liquid in the loop and the water block material. The question however is how much qualitativly is it? "No big difference" is what? The difference of a degree? Half a degree? This is what I'm looking to know.

[airdeano]

no difference on loop order other than reservoir feeds the pump.

on initial thoughts, you'd be correct, but since the system internally will stabilize in temperature, the device order and radiator placement in the loop is a non-issue. basically, shortest tubing runs, aesthetically pleasing, and place the reservoir to feed the water pump.

[johnny5c]

I've had my loop plumbed a few different ways and never noticed any real change in temps. 

[bgibbz]

12 minutes ago, MrWhy said:

Physically and mathmatically speaking there is, and it is affected by a bunch of things not limited to and including the emmisivity of the liquid in the loop and the water block material. The question however is how much qualitativly is it? "No big difference" is what? The difference of a degree? Half a degree? This is what I'm looking to know.

My knoweledge of Chemistry is limited to whatever i learned Freshman year of highschool, but wouldnt temperature remain constant throughout a closed system? Either way, the difference in order should yeild negligible temperature changes. 

[done12many2]

36 minutes ago, MrWhy said:

So I've read things on both sides of the fence and the engineer in me belives the right answer to be that when possible the CPU should always come first (Flow from pump that is) and then the GPU in your loop. But I'm curious on how quantitative the advantage is in doing this, has anyone ever done their own tests on this subject? I'd love to hear your thoughts if you have thanks!

 

I constantly monitor water temps at two locations in my loop.  Just after the GPUs and mid loop after the first three rads in my loop.  Water temps immediately after the GPUs while under load are only 1c higher than water temps after the first three rads.  At idle, temps in those two locations are identical.  A smaller loop may vary a little more, but not by much.  

[xgn]

1 hour ago, MrWhy said:

So I've read things on both sides of the fence and the engineer in me belives the right answer to be that when possible the CPU should always come first (Flow from pump that is) and then the GPU in your loop. But I'm curious on how quantitative the advantage is in doing this, has anyone ever done their own tests on this subject? I'd love to hear your thoughts if you have thanks!

In a car there is a difference but in pc the temperatures is not that high and the flow rates are so high that it reach a equalibrium.

[THFourteen]

5 hours ago, airdeano said:

no difference on loop order other than reservoir feeds the pump.

on initial thoughts, you'd be correct, but since the system internally will stabilize in temperature, the device order and radiator placement in the loop is a non-issue. basically, shortest tubing runs, aesthetically pleasing, and place the reservoir to feed the water pump.

I've read this a lot but seeing as he's asking about the smallest difference, surely water that goes into a radiator will be at a slightly higher temperature than water that comes out of it, as the radiator will have cooled it.

 

Thus it would make sense to feed the water out of the radiator straight into the component that you want to stay at the lowest temperature?

[atomicus]

I don't know what proverbial 'fence' you're talking about... it has been well established and proven that it makes no difference, and in a short time the temp normalizes over the entire loop (with perhaps a 1 degree variance), so don't even worry about it. Just go with whatever loop order looks best and ensure you follow the correct in/out route... that's far more important LOL!

[0ld_Chicken]

Like done12many2, I've got multiple temp sensors. Mine are on either side of all my major heat producing components. 2x 980's and a 4790k (all heavily OC'd) make a maximum of 3°c difference (with pumps on "1 of 5") and exactly 1°c difference with both pumps on 5. That averages to .33° per component, which you won't even be able to read with most monitoring software.

 

Strap that res above the pump and run the loop however you desire!

[xentropa]

GPUs tend to be hotter during gaming, so I would route the water flow so that cold water goes to the CPU Then gpu.  

[atomicus]

12 minutes ago, xentropa said:

GPUs tend to be hotter during gaming, so I would route the water flow so that cold water goes to the CPU Then gpu.  

 

No, it DOES NOT MATTER lol! 

[MrWhy]

7 hours ago, atomicus said:

 

No, it DOES NOT MATTER lol! 

To a point it does matter. While it is true that the "Heat" in the loop will normalize after a time and reach equilibrium at a certain point....this is only a temp equilibrium as far as the loop itself is concerend. If you run the loop with water going through the hottest component first the emmisivity of the water and the way it radiates its heat will be the biggest factor in determing the avg temp of that loop. However I'm thinking if you run the water through the coolest component first the water has time to emit or take in more heat for the next hottest component in effect you would reach two different temperature avgs for each type of loop. In laymans terms I'm aware that the ability of the water to gather and radiate a the speific heat of the system will be negligable but that does not necessarily mean that you couldn't theoretically get an edge taking into account the radiance values of water. But I think I've found my answer. Thanks for your thoughts. I can see I'm digging too far into this, but then again that is what makes it fun!

[atomicus]

20 minutes ago, MrWhy said:

To a point it does matter. While it is true that the "Heat" in the loop will normalize after a time and reach equilibrium at a certain point....this is only a temp equilibrium as far as the loop itself is concerend. If you run the loop with water going through the hottest component first the emmisivity of the water and the way it radiates its heat will be the biggest factor in determing the avg temp of that loop. However I'm thinking if you run the water through the coolest component first the water has time to emit or take in more heat for the next hottest component in effect you would reach two different temperature avgs for each type of loop. In laymans terms I'm aware that the ability of the water to gather and radiate a the speific heat of the system will be negligable but that does not necessarily mean that you couldn't theoretically get an edge taking into account the radiance values of water. But I think I've found my answer. Thanks for your thoughts. I can see I'm digging too far into this, but then again that is what makes it fun!

 

You are definitely overthinking this. I've seen people have multiple temp sensors throughout a loop to test this, and the variance is negligible and fluctuates. In performance terms it literally does not matter. What you're saying may have some scientific basis in fact, but it has no real world benefit to cooling a loop. There is no 'edge'... a degree or two isn't an 'edge' unless you are planning on entering some kind of watercooling competition, in which case you'll have some crazy LN2 set-up anyway. Thermal paste selection and application has a greater impact on temps than any of this, never-mind fans, rads, case etc. Don't worry about it, just go with what looks best and/or is easiest to route, as long as everything is flowing in the right direction. 

[airdeano]

3 hours ago, MrWhy said:

But I think I've found my answer. Thanks for your thoughts. I can see I'm digging too far into this, but then again that is what makes it fun!

see, some of us have already traveled this road and lending experience to save others a lot of expensive time testing, buying parts and equipment, and risk of dousing your rig with a tsunami of coolant.

[MrWhy]

20 hours ago, airdeano said:

see, some of us have already traveled this road and lending experience to save others a lot of expensive time testing, buying parts and equipment, and risk of dousing your rig with a tsunami of coolant.

Haha I do admit this Forum is a great place for info....but from what I'm seeing no. Not many people have traveled the road I'm looking at haha, and what I was looking at doing was not going to douse anything. But it is nice to see so many people willing to lend what experience they have to help a guy out.